Optimal Signal Photon Detection Via Bacterial Foraging Optimization: Maximizing Probability Of Detection In Noisy Environments
DOI:
https://doi.org/10.64252/7gavd784Keywords:
Bacterial Foraging Optimization, Detection Probability, Signal Photons, Evolutionary Algorithms, Likelihood Ratio Test, Photon Detection Model.Abstract
In this work used a Bacterial Foraging Optimization (BFO) algorithm is employed here to solve a photon-counting signal detection task. The MATLAB code implements a bio-inspired algorithm as Bacterial Foraging Optimization (BFO) which is used to maximize the probability of detection (Pd) in signal detection applications. This method emulates bacteria in their search for the optimal signal photon count (S) under noisy conditions where fitness is evaluated by a mathematical likelihood comprising signal photons, background noise (B), coherence factor (M) and maximum photon count threshold (D max). Bacteria adapt and gradually refine their positions in the iterative process until they converge on a best solution. Robust and adaptive even in dynamic environments, it is clear that with processes like chemotaxis, reproduction and elimination-dispersal as iterative the algorithm improves and converges on the best solution. The algorithm presents (Pd) as a function of (S) and graphs how it behaves. Although widely used for optical signal processing applications and remote sensors-to-inspired algorithms are now becoming popular in engineering circles that have been traditionally poorly served by those techniques in its current state of the art. This code must stand as a testament, that by drawing on life for inspiration, bringing together math and nature to address these difficult problems involving complex optimization it offers an utterly trustworthy way of raising both your detection power and its reliability.to numerically address a photon-counting signal detection problem with B equal 46 background noise We adopted a population size of N=10 bacteria that iteratively explores the search-space S ∈[74,200] using chemotactic motion (step-size C=0.1), swims towards higher fitness regions (Ns=4), reproduces the fitter half [N re =4], and eliminates or disperses (Ned=2,ped=25%) by locality to evade local optima. Detection probability is ascertained using a Likelihood Ratio Test (LRT) extracted from Poisson photon statistics. Here Pd specifies the probability of S photons being detected against noise, while
D thresh is found by comparing hypothesis H1 (signal-plus-noise) against H0 (noise alone).
